IoTChat

Big Data Becomes Fast Data
Wikibon Research reviews latest research about how apps are changing as #IoT products generate data
   9 years ago
#IoTChatIoT & Hyperconverged AnalyticsWikibon Analysts hosing data community on topic of emerging era of IoT and hyper-converged analytics
John Furrier
Q: What are your final thoughts for #IoT? Open thread. Share your vision or comments
George Gilbert
every device becomes a part of a larger ecosystem
George Gilbert
every device becomes potentially as smart as the entire cloud that's behind it
John Furrier
I'm loving this trend for global #iot for all devices and people; the thing I think the most about is the developer role and open source.
John Furrier
and the definition of what I call the "True Edge of the Network"
George Gilbert
that's the biggest open topic: how people will program apps built on combination of #fastdata and #bigdata
George Gilbert
#opensource pricing becomes even more critical because data volumes grow by several orders of magnitude beyond today's #bigdata
Muddu Sudhakar
Consumer IOT and Industrial IOT are separate use cases and markets
Manickam Anandan
when d connectivity on d places existed the iot growth ll increase... The application made fa health purpose d reach of iot ll increase...
Muddu Sudhakar
it is not clear if there is good business model for Consumer IOT market and who/what way can vendors drive revenue
Sanjeev Kumar
@ggilbert41 some leading indicators are - CEP / stream programming model, sliding time windows, stats based analytics and models
mike_davis
Insurance gets cheaper, fitness improves, things break down less often, customer support gets less dumb. In short no one will notice!
Dave Vellante
The big industrial giants are not working together today (e.g. GE, IBM, Hitachi...) and they will realize to advance the market they must collaborate
Manickam Anandan
finally wind up wit tis...iot s the future world Bt wen tat is mingled wit people like daytoday life things... It get much fast to reach the people...
George Gilbert
@smuddu what about B2B2C model - like wearables that feed medical devices and healthcare providers and then consumer advice
Muddu Sudhakar
Fitbit is in good position to monetize consumer IOT for healthcare, insurance markets. The market has not evolved yet in this space to drive new business models and help consumers
Sanjeev Kumar
@ggilbert41 yes IoTs would mainly be a B2B2C or B2B2B model
George Gilbert
@smuddu @dataplumbers insurance underwriters can become providers of b2c prescriptive analytics by, for example, providing diet advice or route guidance when trying to avoid bad neighborhoods
Crowd Captain
Ron Bodkin ‏@ronbodkin said to @furrier #IoT value will be in the analytics - directly observing what's happening will go from behavioral insights to data science action
Muddu Sudhakar
@dvellante good that big industrial giants are NOT working together otherwise this will slowdown adoption and creates unnecessary standards. IOT is in early days and the best way new products/solutions delivered is by focusing on problem/use case
Sanjeev Kumar
@smuddu at the same time you can get "walled gardens" supported by GE, 3M ... which might make the life of their customers quite miserable
John Furrier
Q: What is the fantasy of #IoT that is the future but not yet doable today for CxOs?
Ryan Begley
self driving cars at scale - highly efficient traffic
Sanjeev Kumar
programmable sensors connected to edge networks connected to cloud backends hosting ML software to generate prediction models
Sanjeev Kumar
to complete the above - and the prediction models adjust / tune the sensors dynamically
John Furrier
To me we've seen this movie before with "last mile" internet and wireless broadband. Truck rolls (provisioning) will kill the economic models except for the totally obvious game-changing use cases
Ryan Begley
I am also looking for connected care, elder care
John Furrier
Automating away the people equation for getting the real time data is the key
George Gilbert
game-changing use cases are starting with heavy industrial devices - they will have power and may have intermittent connectivity
Michael G. Nelson
What happens when the privacy pendulum swings hard to the right?
Crowd Captain
Many will miss the boat on #iot bc they try to many outside the edge of network use cases and cost overruns will kill them
George Gilbert
the connectivity problem is being solved by using buffer processing that stores data periodically captured from a device like a jet engine, car, windmill
Keith Townsend - Light will overcome darkness
IoT security I think is a big scare for CxO's. It's complex enough to secure systems within a data center. IoT devices are a security nightmare.
John Furrier
@abusedbits Privacy is huge and #iot will open up the surface area for attacks so security has to be built in NOT bolted on
Michael G. Nelson
@ggilbert41 so it's not Fast Data, it's periodically fast data? :-)
John Furrier
@abusedbits data can trickle to a new notion of "smart edge router/server/dataprocessor
George Gilbert
surface area is huge. but stream processors can also leverage encryption with keys from security manager
George Gilbert
smart edge is what makes it possible to handle occasionally connected devices
Michael G. Nelson
So, the future is an NFV integration at the edge for services consolidation for the data. Can't wait to see that compete for the same space as video streaming....
George Gilbert
@abusedbits periodically fast data is great term
Sanjeev Kumar
Home IoT fantasy - you walk into the house, thermostat has been set a few mins ago, kitchen appliances go on standby, TV/ent displays a summary of new content, music play in the background
George Gilbert
@abusedbits periodically fast is crucial also because of its implication for #IoT programming model
Crowd Captain
Android Developer ‏@androider said to @furrier
it just so happens I'm working on this. So yet it will be available soon.
Sanjeev Kumar
Home IoT fantasy cont: of course you came home in a self-driven car, that communicated to the IoT hub at home to trigger the downstream activities
David Floyer
The starting point is the cost of sensor & sensor deployment. Where sensors are derived from consumer technologies (e.g., mobile) the costs can be low. Edge computing has to be very low cost and integrated (like consumer set-top boxes)
Crowd Captain
@abusedbits fast data was coined by @furrier years ago
Sanjeev Kumar
@dfloyer cost per sensor is now down to almost $1
Michael G. Nelson
coined periodically fast data or PFD if you will... @CrowdCaptain
Smarty
Self Driving, Levitating cars #Jetsons